A fuzzy BP approach for diagnosis and prognosis of bearing faults in induction motors

This paper demonstrates a novel and cost-effective approach for diagnosis and prognosis of bearing faults in small and medium size induction motors. Even though, many researchers dealt with the bearing fault diagnosis of induction motors by using traditional and soft computing approaches, the application of these techniques for predicting the remaining life time of electrical equipment is not seen much in the literature. Moreover, individual artificial intelligence (AI) techniques suffer from their own drawbacks, which can overcome by forming a hybrid approach combining the advantages of each technique. Hence, in this paper an attempt has been made to combine neural networks and fuzzy logic and forming a fuzzy back propagation (fuzzy BP) network for identifying the present condition of the bearing and estimate the remaining useful time of the motor. The results obtained from fuzzy BP network are compared with the neural network, which show that the hybrid approach is well suitable for assessing the present condition of the bearing and the time available for the replacement of the bearing.